Literature Review

A recent report on Purdue University’s campus life said that vegetarians are facing limited meal options in dining halls. Vegetarians complain that most of the times, they only have tofu and not too many healthy vegetarian options. Studies also prove that on average, in the US, meat is consumed more than three times the global average, and in much more quantity than vegetarian food. This proves that the American society is a big consuming society; they believe that bigger is better and value quantity over quality. It is therefore not surprising that many organizations and academic institutions may follow this belief and tend to care less about other cultures’ ways of eating.

Another article highlights the ‘American Ideal Dinner’ and how it focuses on red and processed meat, thereby normalizing red meat options at university dining spaces and their overcompensation for fresher, more diverse vegetarian options. Being an international student and having suffered a major shift in weight and physical health in my freshman year at Emory, I was interested in learning if other students felt the same way or if they were satisfied with the current dining facilities. I also wanted to understand whether sex played a role in maintaining a particular dietary preference (omnivorous or non-omnivorous) but the same Foodways in Focus article concludes that there is only a 1% difference in the number of women who choose the vegetarian lifestyle and men who do the same.

With this information, I was motivated to survey Emory University students and find out if students with different dietary needs, have any complaints regarding the meal options at DCT, Cox Hall, and other on-campus dining spots, and if they resort to home cooking due to lack of options or other personal health-related reasons.

Balanced Diet

Hypothesis

Are non-omnivorous students at Emory unhappy with the limited meal options, compared to omnivores, and has it played a role in impacting their overall health and happiness levels?

Null Hypothesis (H0) : Omnivores have higher Average Happiness Levels at Emory, than non-Omnivores.

Explanation of Groups

Using the dplyr package and the rename function, I changed the column names as follows: q12 - Happiness q13 - Diet (for Dietary Preferences) q33 - Cook_freq (for how frequently students cook on their own) q34 - Pre_made (for how frequently students order pre-made food)

Following this, I cleaned the Dataset to present the table in a neater fashion.

Understanding Dietary Preferences

The following tables quantify the number of students that belong to each diet category, and what their sex and ethnicity is.

## 
##     Kosher   Omnivore      Other      Vegan Vegetarian 
##          2        116          2          3         15

## 
##     Kosher   Omnivore      Other      Vegan Vegetarian 
## 0.01449275 0.84057971 0.01449275 0.02173913 0.10869565

The graph below will analyze whether these dietary preferences have any relationship with their average overall happiness levels at Emory so far.

Measuring Happiness

## # A tibble: 9 x 3
## # Groups:   sex, Diet [9]
##   sex    Diet       AvgHappiness
##   <chr>  <chr>             <dbl>
## 1 Female Omnivore           3.46
## 2 Female Other              4   
## 3 Female Vegan              3   
## 4 Female Vegetarian         3.69
## 5 Male   Kosher             3.5 
## 6 Male   Omnivore           3.75
## 7 Male   Other              4   
## 8 Male   Vegan              3.5 
## 9 Male   Vegetarian         3.79

## # A tibble: 5 x 2
## # Groups:   Diet [5]
##   Diet       AvgHappiness
##   <chr>             <dbl>
## 1 Kosher             3.5 
## 2 Omnivore           3.66
## 3 Other              4   
## 4 Vegan              3.33
## 5 Vegetarian         3.73

Observing Ethnicity

Scatterplot 2 below depicts Diet against Ethnicity and Happiness, and has not taken sex into consideration because the purpose of this graph is to identify if people of certain Ethnicities respond to meal options in a particular way, such that it affects their Happiness levels.

## # A tibble: 138 x 7
## # Groups:   Ethnicity, Diet, Happiness [37]
##    Happiness Diet       Cook_freq Pre_made Ethnicity      sex    control
##        <dbl> <chr>          <dbl>    <dbl> <chr>          <chr>  <fct>  
##  1         3 Vegetarian        14        3 American       Male   Male   
##  2         2 Omnivore           2        8 Asian          Male   Male   
##  3         4 Omnivore           1        1 Asian          Male   Male   
##  4         1 Omnivore           4        3 Asian          Male   Male   
##  5         4 Omnivore           7        0 American       Male   Male   
##  6         4 Omnivore           3        2 Asian          Male   Male   
##  7         3 Omnivore           2        1 Asian          Male   Male   
##  8         4 Other              7        1 Asian          Female Female 
##  9         3 Omnivore           7        3 Middle Eastern Male   Male   
## 10         3 Omnivore           0        2 Asian          Male   Male   
## # … with 128 more rows

According to this Scatterplot above, Vegan Americans seem rather happy with their life at Emory (Happiness levels 3-4), which is interesting because this again proves that they do not feel that food plays a negative role in their lives. Omnivores, mostly belong to the Asian category, and vary in their happiness levels from 1-5. This could be due to personal tastes and preferences, and behavioral attitudes towards food. Asians in general, seem to be more on the happier end (except for one outlier), and this reflects upon the abundance of asian food options available across campus, at the DCT, Cox Hall, Rollins, etc. Kosher Hispanics, though taking up a small portion of the sample population, are neutral (happiness level 3). It is understood that Kosher meals are not as widespread or available as other options.

The above column plot links ethnicity and frequency of consuming pre-made food, which is not supposed to be the healthiest option. Results show that Asians are most active in ordering pre-made food, and it would be interesting to find out what kinds of pre-made food are they purchasing - frozen meals, restaurant-made meals, easy-to-bake meals, etc. This would provide more information on their overall health status.

Observing Cooking Frequencies of Students

Cooking Frequencies have been numbered in ‘times per week’. Using kable to design this beautiful table, I have been able to define how many students cook more frequently in a week, compared to others. For instance, 25 students never cook at home. 19 students cook approximately once a week, 4 students cook almost every day, etc.

## 
##  0  1  2  3  4  5  6  7  8 10 12 14 15 16 21 
## 25 19 18 16 11 17  3 13  2  7  1  2  2  1  1

## 
##           0           1           2           3           4           5 
## 0.181159420 0.137681159 0.130434783 0.115942029 0.079710145 0.123188406 
##           6           7           8          10          12          14 
## 0.021739130 0.094202899 0.014492754 0.050724638 0.007246377 0.014492754 
##          15          16          21 
## 0.014492754 0.007246377 0.007246377

The graph above shows the relationship between how happy a student is, and how frequently they cook on their own as a substitute to dining on campus. Students who cook approximately 7 times a week, are the happiest. This means they cook rather frequently, perhaps one of three meals every day. Those who cook the most (more than 16 times a week) have low levels of happiness and one can assume that the added stress and effort of cooking everyday takes a toll on the happiness levels of a student.

This violin plot graphs a relationship between sex of a student and how often they cook. Traditional beliefs have told us that females cook more than males, and while this might be true for a small portion of the female violin plot represented by the tip of the graph that extends just beyond 15 times/week, it is obvious that the densest portion of both plots is towards the lower end of cooking frequencies implying that the majority of sample population does not cook (regardless of the sex)

Analysis of Confidence Intervals and Testing

A T-test is conducted to determine the significance of the difference in means observed for each group being examined. For this study, I have analyzed the 5 categorical groups of Dietary Preferences (Omnivorous, Vegetarian, Vegan, Kosher, Other).

The two T-tests show the relationship between Diet types and Average Happiness, on two groups based on the control (i.e. Sex) e.g. Male Omnivores and Male Vegetarians, and Female Omnivores and Female Vegetarians.

This confirms whether there is a statistical difference between the mean level of Happiness between the two sets of groups.

The level of significance for this test is an alpha of 0.05.

## `summarise()` regrouping output by 'Diet' (override with `.groups` argument)

## 
##  Welch Two Sample t-test
## 
## data:  Happiness by Diet
## t = -0.11506, df = 7.8893, p-value = 0.9113
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.6864758  0.6213764
## sample estimates:
##   mean in group Omnivore mean in group Vegetarian 
##                 3.753165                 3.785714

## 
##  Welch Two Sample t-test
## 
## data:  Happiness by Diet
## t = -0.66321, df = 14.703, p-value = 0.5175
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.9622219  0.5061408
## sample estimates:
##   mean in group Omnivore mean in group Vegetarian 
##                 3.459459                 3.687500

Based on our findings…

This project set out with the intention of exploring the relationship between Student Diet Types and Happiness with the overall Emory Community, to confirm whether the two were at all related and if Emory meal options played a role in determining the happiness levels of students.

The two hypotheses leading this investigation were that (1) students who consumed meat (omnivores) had a higher average happiness level than students with other dietary restrictions (2) there was a statistical difference between mean Happiness levels of Omnivores and Vegetarians (the two most common dietary preferences of students).

The first hypothesis was based on the prior literature discussed, that the Ideal American Dinner comprises red and processed meat and therefore it is expected that most institutions will favor an omnivorous selection of food rather than vegan, vegetarian and kosher options. Additionally, students from other universities have also complained that there are not enough vegetarian options on campus to suit their diet needs.If this were to be true, then non-omnivores should be less satisfied with the Emory Community and show display lower levels of Average Happiness.

The second hypothesis was guided by a similar idea that there might be a significant statistical difference between the happiness levels of the two groups.

The results of this study, however, rejected both the null hypotheses. There are a variety of factors that could lead into why this is the case. Firstly, the sample selected was limited to Emory University undergraduate students who were enrolled in ECON220L: Probability & Statistics for Economists Lab. We cannot be sure that the rest of the student population feels the same way about dining options and restrictions. The second reason is tied to the first, in that, since we had a tight sample group, the number of omnivores was significantly higher than the number of students with the other 4 dietary preferences. For this reason, the overall raw results would be biased by this huge number of Omnivores. Given a larger sample of the whole student body, there might be more vegans and vegetarians to balance the data out. Third, since our ‘Happiness’ variable speaks of overall happiness at the University, it is possible that Diet and on-campus meal options do not play a significant role in determining how happy a student is at the university. Other factors such as education, faculty, cleanliness and social circles might be heavier contributors to their happiness levels. The fourth reason (a possibly minor reason) is that since this study was based on a self-report method of data collection, students might have inaccurately documented how often they cook at home or order pre-made food. This could have led to slightly skewed results, though it may not be the prime reason why.

Further Applications

Despite the two hypotheses lacking support, this project was still valuable in various ways. Firstly, it informed me against my belief that omnivores are more satisfied with life at Emory whereas Vegetarians and Vegans feel more restricted and bound by unhealthy options, adversely affecting their well-being. Looking forward, it would be interesting to explore Happiness level that is specifically connected to food at Emory. Students could respond to surveys about their satisfaction with the new DCT compared to the previous years’ DUC and rate their happiness levels, health levels, etc. It would then be a worthy pursuit to use similar variables of Cooking Frequency and Pre_made Frequency to understand if Omnivores are really living a healthier lifestyle than Non-Omnivores.

Studies like these would go a long way in improving the overall standard of living at Emory, and take a dive into the student body’s perspectives of the university’s upkeep.